Loading/unloading data occupancy dataset - Barcelona LL - DISCO project

DISCO Project Dataset: Computer Vision Parking Detection and Monitoring Overview This dataset contains computer vision-based vehicle detection, tracking, and parking monitoring data collected in Barcelona un/loading areas during December 2025 - March 2026. The project aimed to provide comprehensive vehicle occupancy monitoring that complements the official [SPRO parking management system](https://webspro.bsmsa.eu/), which records only users who actively check in via the mobile application. The dataset covers two distinct monitoring areas with different camera setups and collection methodologies, enabling researchers to study vehicle detection, license plate recognition, parking occupancy patterns, and dwell time analysis in real-world urban environments. Study Areas Area 4703: Pedestrian Street (Multicamera Setup) A pedestrian street with 10 parking spaces reserved for un/loading operations during specific hours. Location: CONSELL DE CENT, 378 Operational hours**: 07:00 - 08:00 and 10:00 - 16:00 (reserved for un/loading) Recording period**: 06:30 - 16:30 Monitored using two fixed cameras: Camera 2: Entry point monitoring Camera 3: Exit point monitoring Data is collected continuously with frame-based processing. License plates detected at entry and exit are matched using OCR-read plate numbers to link entrance and exit records, creating complete tracking of individual vehicles through the parking area. Area 3177: Chamfered Area (Single Camera Setup) A chamfered parking area with multiple parking slots reserved for un/loading operations during specific hours. Location: VALENCIA, 394 Operational hours: 08:00 - 20:00 (reserved for un/loading) Recording period**: 07:30 - 20:30 Monitored using a single fixed camera with frames collected at 30-second intervals (not continuous), providing periodic snapshots of the parking situation. This setup captures long-term occupancy patterns and is suitable for areas where continuous video monitoring is not feasible. For more information, see the README.md file in the dataset.